`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`The New England Journal of Medicine
`
`COMBINATION THERAPY WITH EFAVIRENZ, NELFINAVIR, AND NUCLEOSIDE
`REVERSE-TRANSCRIPTASE INHIBITORS IN CHILDREN INFECTED WITH
`HUMAN IMMUNODEFICIENCY VIRUS TYPE 1
`
`, M.S.,
` H. Y
`, M.D., F
` A. S
`.D., S
`, P
` V. F
`, M.D., C
` E. S
`S
`ONG
`LORENCE
`PECTOR
`TEPHEN
`HARM
`LETCHER
`OURTNEY
`TARR
`TUART
`T
` F
`, E
`.D., R
` C. B
`, P
`.D., P
`.D., D
` M
`, M.D., N
` M. R
`, M.D.,
`H
`HARM
`RUNDAGE
`D
`ICHARD
`ERENCE
`ENTON
`OUGLAS
`ANION
`ANCY
`UIZ
`M
` G
`, M.D., M
` B
`, P
`.D., J
` M
`N
`, M.D.,
` L
` M. M
`, M.D.,
`ERRIL
`ERSTEN
`ARK
`ECKER
`HARM
`AMES
`C
`AMARA
`AND
`YNNE
`OFENSON
`
` P
` AIDS C
` T
` G
` 382 T
`*
`FOR
`THE
`EDIATRIC
`LINICAL
`RIALS
`ROUP
`EAM
`
`A
`BSTRACT
`Background
`Consistent long-term viral suppres-
`sion has been difficult to achieve in children with hu-
`man immunodeficiency virus type 1 (HIV-1) infection.
`We tested the safety and antiviral efficacy of a novel
`combination consisting of efavirenz, nelfinavir, and
`one or more nucleoside reverse-transcriptase inhib-
`itors in 57 children previously treated with only nu-
`cleoside reverse-transcriptase inhibitors.
`Methods
`The children were monitored for 48 weeks
`after the initiation of therapy. We assessed plasma
`concentrations of efavirenz and nelfinavir, plasma
`HIV-1 RNA levels, and lymphocyte subpopulations.
`Results
`At base line, the 57 HIV-1—infected children
`(age range, 3.8 to 16.8 years) had a median of 699
`CD4 cells per cubic millimeter and 10,000 copies of
`HIV-1 RNA per milliliter of plasma. The most common
`treatment-related effects of at least moderate sever-
`ity were rash (in 30 percent of children), diarrhea (in
`18 percent), neutropenia (in 12 percent), and bio-
`chemical abnormalities (in 12 percent). Serious side
`effects were uncommon. The mean values for the
`area under the curve for efavirenz and nelfinavir cor-
`responded to expected values. In an intention-to-treat
`analysis, 76 percent of children had plasma HIV-1
`RNA levels of less than 400 copies per milliliter after
`48 weeks of therapy and 63 percent had levels of less
`than 50 copies per milliliter. A high plasma HIV-1 RNA
`level at base line significantly decreased the likeli-
`hood that plasma levels of HIV-1 RNA would become
`undetectable during treatment.
`Conclusions
`In HIV-1—infected children who were
`previously treated with nucleoside reverse-transcrip-
`tase inhibitors, the combination of efavirenz, nelfina-
`vir, and nucleoside reverse-transcriptase inhibitors
`was generally well tolerated and had a potent and sus-
`tained antiviral effect. (N Engl J Med 1999;341:1874-81.)
`'1999, Massachusetts Medical Society.
`
`A
`
` MAJOR objective of antiretroviral ther-
`apy is to reduce the amount of human im-
`munodeficiency virus type 1 (HIV-1) RNA
`in plasma to undetectable levels, because
`these levels appear to reflect the degree of viral rep-
`lication in the body. Even low levels of replication may
`contribute to the emergence of resistant strains of
`HIV-1. Studies of highly active antiretroviral therapy
`have generally yielded less impressive results in HIV-
`infected children than in infected adults. In small stud-
`
`1874
`
`Æ
`
`December 16, 1999
`
`ies of combinations that included a protease inhibitor
`in children who had previously been treated with nu-
`cleoside reverse-transcriptase inhibitors, plasma HIV-1
`RNA levels decreased to undetectable levels (less than
`400 copies per milliliter) in 39 to 78 percent, but
`subsequently rebounded in 20 to 44 percent.
` In an-
`1-5
`other study, only two of eight HIV-1—infected infants
`had undetectable plasma HIV-1 RNA levels after com-
`bination therapy with zidovudine, didanosine, and ne-
`virapine.
` Clearly, other promising antiretroviral com-
`6
`binations need to be tested in children.
`Efavirenz, a new nonnucleoside reverse-transcrip-
`tase inhibitor, has good oral bioavailability and a long
`terminal half-life, permitting once-daily dosing.
` We
`7,8
`tested efavirenz in combination with a protease in-
`hibitor in children, the majority of whom were re-
`ceiving one or more nucleoside reverse-transcriptase
`inhibitors, in accordance with recommendations that
`at least two new antiretroviral drugs be added to an
`existing regimen.
` Nelfinavir was selected as the pro-
`9
`tease inhibitor because it is available in a pediatric
`formulation, has approved dosing recommendations
`for children older than two years of age, and has no
`pharmacokinetic interactions with efavirenz.
`10
`
`METHODS
`
`Study Design
`This open-label study was conducted at 18 Pediatric AIDS
`Clinical Trials Group (PACTG) sites. Eligibility criteria included
`
`From Children(cid:213)s Hospital of Philadelphia, Philadelphia (S.E.S.); the Uni-
`versity of Minnesota, Minneapolis (C.V.F., R.C.B.); the University of Cal-
`ifornia at San Diego, San Diego (S.A.S.); Harvard School of Public Health,
`Boston (F.H.Y.); Frontier Science and Technology Research Foundation,
`Brookline, Mass. (T.F.); Dupont Pharmaceuticals Company, Wilmington,
`Del. (D.M., N.M.R.); Agouron Pharmaceuticals, La Jolla, Calif. (M.G.,
`M.B.); the National Institute of Allergy and Infectious Diseases, Rockville,
`Md. (J.M.); and the National Institute of Child Health and Human Devel-
`opment, Rockville, Md. (L.M.M.). Address reprint requests to Dr. Starr at
`the Division of Immunologic and Infectious Diseases, Children(cid:213)s Hospital
`of Philadelphia, 34th St. and Civic Center Blvd., Philadelphia, PA 19104,
`or at starr@email.chop.edu.
`Other authors were Lynette Purdue, Pharm.D., National Institute of Al-
`lergy and Infectious Diseases, Rockville, Md.; Suzanne Siminski, M.S., and
`Bobbie Graham, B.S., Frontier Science and Technology Research Founda-
`tion, Amherst, N.Y.; David M. Kornhauser, M.D., and William Fiske, Ph.D.,
`Dupont Pharmaceuticals Company, Wilmington, Del.; Carol Vincent,
`C.R.N.P., Children(cid:213)s Hospital of Philadelphia, Philadelphia; Harold W.
`Lischner, M.D., St. Christopher(cid:213)s Hospital for Children, Philadelphia;
`Wayne M. Dankner, M.D., University of California at San Diego, San Di-
`ego; and Patricia M. Flynn, M.D., St. Jude Hospital for Children, Mem-
`phis, Tenn.
`*Other participants in the study are listed in the Appendix.
`
`Page 1 of 8
`
`Biogen Exhibit 2055
`Coalition v. Biogen
`IPR2015-01993
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`COMBINATION THERAPY IN CHILDREN INFECTED WITH HIV-1
`
`an age of less than 16 years, a plasma HIV-1 RNA level of more
`than 400 copies per milliliter on measurement with a quantitative
`RNA reverse-transcriptase—polymerase chain reaction assay (Am-
`plicor HIV-1 Monitor assay, Roche Diagnostic Systems, Branch-
`burg, N.J.), no previous treatment with nonnucleoside reverse-trans-
`criptase inhibitors or protease inhibitors, concomitant treatment
`with at least one nucleoside reverse-transcriptase inhibitor, and
`the ability to swallow capsules. The children were enrolled be-
`tween October 27, 1997, and February 12, 1998.
`Efavirenz (Sustiva, Dupont Pharmaceuticals, Wilmington, Del.)
`was provided by the manufacturer in 50-mg, 75-mg, and 100-mg
`capsules. Nelfinavir (Viracept, Agouron Pharmaceuticals, La Jol-
`la, Calif.) was provided by the manufacturer in tablet (250 mg)
`and powder (50 mg per gram) form. The initial doses of efavirenz
`were allometrically scaled to body size according to the following
`formula: the dose (in milligrams per day)=(the weight of the child
`in kilograms(cid:214)70)
`´600. The doses were rounded to the near-
`0.7
`est 25-mg increment. Efavirenz was given daily in the morning.
`The initial dose of nelfinavir was the recommended pediatric dose
`of 20 to 30 mg per kilogram of body weight thrice daily. At en-
`try, the children continued to take the same nucleoside reverse-
`transcriptase inhibitors or were switched to new ones, at the dis-
`cretion of the site investigators.
`The doses of efavirenz and nelfinavir were adjusted if the area un-
`der the curve was too small. The target value for the area under the
`curve from 0 to 24 hours was 190 to 380 (cid:181)mol per liter ?
`hour for
`efavirenz, representing the range from the 50th percentile to twice
`the 50th percentile of such values in adults receiving 600 mg per
`day of efavirenz. This target value was selected to help ensure that
`therapeutic levels would be achieved in this pediatric study. For nel-
`finavir, the target value for the area under the curve from zero to
`eight hours was 10 mg per liter ?
`hour, the 10th percentile for adults
`receiving 750 mg of nelfinavir thrice daily. The areas under the
`curve for efavirenz and nelfinavir were determined at weeks 2 and
`6 and again at week 10, if necessary. Blood samples were obtained
`before a dose and 2, 5, 6, 8, 12, and 24 hours after a dose. If target
`values for the area under the curve were not achieved, the doses of
`efavirenz, nelfinavir, or both were adjusted proportionally.
`History taking and physical examinations were performed at
`base line; at weeks 2, 4, 6, and 8; and every four weeks thereafter.
`A complete blood count and differential count were obtained and
`blood chemical analysis and urinalysis were performed at base line
`and at weeks 2, 4, 8, 12, 16, 20, 24, 32, 40, and 48. Specimens
`for the determination of plasma HIV-1 RNA levels and lympho-
`cyte subpopulations were obtained at base line and at weeks 2, 4,
`8, 12, 20, 32, 40, and 48.
`We used the Division of AIDS Toxicity Table to grade the se-
`verity of adverse effects. A grade of 0 indicates the absence of ad-
`verse effects; a grade of 1 indicates mild adverse effects, a grade
`of 2 moderate effects, a grade of 3 severe effects, and a grade of
`4 life-threatening effects. Treatment was discontinued if a child
`had grade 4 effects, repeated grade 3 effects, or poor compliance,
`defined as consumption of less than 80 percent of the doses of
`efavirenz or nelfinavir over a one-month period. Virologic failure
`was defined as a reduction in plasma HIV-1 RNA levels of less
`than 1 log (on a logarithmic [base 10] scale) by week 12, HIV-1
`RNA levels of more than 1000 copies per milliliter on two suc-
`cessive measurements in children whose levels had been less than
`400 copies per milliliter, or HIV-1 RNA levels that were more
`than 0.75 log copies above the nadir value on two successive oc-
`casions in children whose plasma HIV-1 RNA levels had remained
`detectable.
`The institutional review boards at each site approved the study.
`Written informed consent was obtained from the parents or guard-
`ians of all children.
`Measurement of Plasma Efavirenz and Nelfinavir
`Concentrations
`Concentrations of efavirenz and nelfinavir in plasma were meas-
`ured by high-performance liquid chromatography. The total vari-
`ability (expressed as a composite of within-day and day-to-day vari-
`
`ability) of the efavirenz and nelfinavir assays was 1 to 4.5 percent
`and 1.7 to 10.3 percent, respectively.
`
`Determination of Lymphocyte Subpopulations
`and Quantitation of Plasma HIV-1 RNA
`The percentages and absolute numbers of CD4 cells were de-
`termined in PACTG-certified laboratories according to the PACTG
`consensus protocol for flow cytometry.
` Plasma was stored at
`11
``70¡C, and HIV-1 RNA was measured with a quantitative assay
`(Amplicor, Roche Diagnostic) with a lower limit of quantitation
`of 400 copies per milliliter. Samples containing less than 400 cop-
`ies per milliliter were tested with an ultrasensitive assay with a
`lower limit of quantitation of 50 copies per milliliter. All RNA as-
`says were performed in a single PACTG-certified laboratory.
`
`Statistical Analysis
`Primary efficacy was assessed on the basis of plasma HIV-1 RNA
`levels. For each child, a series of composite HIV-1 RNA measures
`was constructed as follows: the result of the quantitative assay was
`used, unless it was less than 400 copies per milliliter or was not
`available, in which case the result of the ultrasensitive assay was
`used. Log-transformed values were used for analyses.
`Three types of analysis were used to evaluate primary efficacy: an
`analysis according to the intention to treat, in which all HIV-1 RNA
`levels were included, regardless of whether a child was still receiv-
`ing drugs; an analysis in which only the HIV-1 RNA levels of chil-
`dren who were receiving treatment were included; and an analysis
`in which HIV-1 RNA levels were assigned a value of more than 400
`copies per milliliter for all visits after the discontinuation of treat-
`ment (or a value of more than 50 copies per milliliter with the use
`of the ultrasensitive assay). Missing HIV-1 RNA values were as-
`signed as a value of more than 400 copies per milliliter (or more
`than 50 copies per milliliter) if the first value preceding or follow-
`ing the missing value exceeded 400 (or 50) copies per milliliter. No
`value was imputed for a missing result if preceding and succeeding
`values were both less than 400 or less than 50 copies per milliliter.
`We calculated exact 95 percent confidence intervals for the per-
`centage of children with a virologic response at each visit, assum-
`ing a binomial distribution. We determined the median change in
`plasma HIV-1 RNA levels from base line and corresponding 95
`percent confidence intervals according to the method of Brook-
`meyer and Crowley, in which HIV-1 RNA levels below the limit
`of detection were treated as censored observations.
`12
`All other analyses used an intention-to-treat approach unless
`otherwise noted. The times to various outcomes were estimated
`according to the Kaplan—Meier method.
` We used the Cox pro-
`13
`portional-hazards regression model to estimate relative risks and
`corresponding 95 percent confidence intervals.
` We used a step-
`14
`wise proportional-hazards model to assess the relative importance
`of risk factors in predicting the likelihood that plasma HIV-1
`RNA levels would become undetectable. Weight measurements
`were converted to z scores after adjustment for age and sex.
` The
`15
`effect of treatment on CD4 measures was examined with use of
`the Wilcoxon signed-rank test (pairwise comparison), and the one-
`sample application of the Wei—Johnson test, in which components
`of the global test statistic were weighted by reciprocals of variances
`of the test statistic at each time.
` The 95 percent confidence in-
`16
`tervals for the median changes in the absolute numbers and per-
`centages of CD4 cells were based on order statistics.
` The Wilcox-
`17
`on rank-sum test was used to compare the duration of rash between
`children.
` All reported P values are two-sided.
`18
`
`RESULTS
`Base-Line Characteristics
`The base-line characteristics of the 57 children who
`were enrolled in the study are shown in Table 1. Be-
`fore enrollment, 55 children were receiving one or
`more nucleoside reverse-transcriptase inhibitors. Thir-
`
`Volume 341 Number 25
`
`Æ
`
`1875
`
`Page 2 of 8
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`The New England Journal of Medicine
`
`ty-four children (60 percent), including 2 who were
`not already receiving nucleoside reverse-transcrip-
`tase inhibitors, began to receive at least one new nu-
`cleoside reverse-transcriptase inhibitor within two
`weeks after study entry, 21 (37 percent) had no
`change in their regimen of nucleoside reverse-trans-
`criptase inhibitors, and 2 (4 percent) had one nucle-
`oside reverse-transcriptase inhibitor dropped from
`their regimen.
`
`Adverse Effects
`Twenty-six children (46 percent) had either no
`adverse effects or no more than mild effects, 25 (44
`percent) had no more than moderate effects, 5 (9 per-
`
`T
`
`ABLE
`
` C
`-L
` B
` 1.
`HARACTERISTICS
`INE
`ASE
`57 C
`
`
` S
`.
`IN
`THE
`TUDY
`HILDREN
`
`
`
`OF
`
`
`
`THE
`
`
`
`C
`HARACTERISTIC
`
`Sex (cid:209) no. (%)
`Female
`Male
`Race or ethnic group (cid:209) no. (%)
`Non-Hispanic white
`Non-Hispanic black
`Hispanic (all races)
`HIV risk factors (cid:209) no. (%)
`Maternal transmission
`Transfusion before July 1985
`Sexual partner with HIV
`Sexual abuse
`CDC class (cid:209) no. (%)*
`Class N (asymptomatic)
`Class A (mild)
`Class B (moderate)
`Class C (severe)
`Age (cid:209) yr
`Mean –SD
`Median
`Range
`CD4 cell count (cid:209) per mm
`3
`Mean –SD
`Median
`Range
`CD4 cell count (cid:209) %
`Mean –SD
`Median
`Range
`Weight (cid:209) kg
`Mean –SD
`Median
`Range
`Weight (cid:209) z score(cid:160)
`Mean –SD
`Median
`Range
`Plasma HIV-1 RNA level (cid:209) log copies/ml(cid:224)
`Mean –SD
`Median
`Range
`
`V
`ALUE
`
`37 (65)
`20 (35)
`
`9 (16)
`33 (58)
`15 (26)
`
`52 (91)
`2 (4)
`2 (4)
`1 (2)
`
`4 (7)
`32 (56)
`18 (32)
`3 (5)
`
`8.5–3.3
`8.0
`3.8 to 16.8
`
`841.1–513.8
`699
`4 to 2616
`
`30.6–10.7
`30
`1 to 56
`
`27.9–14 2
`24.8
`13.2 to 96.4
`
``0.25–1.11
``0.34
``2.47 to +3.21
`
`4.1–0.7
`4.0
`2.6 to 5.7
`
`*CDC denotes Centers for Disease Control and Preven-
`tion.
`(cid:160)The z scores were adjusted for age and sex.
`(cid:224)The quantitative assay was used. Two subjects had HIV-1
`RNA levels of less than 400 copies per milliliter at entry, one
`of whom had an HIV-1 RNA level of 1180 copies per milli-
`liter in the ultrasensitive assay.
`
`1876
`
`Æ
`
`December 16, 1999
`
`cent) had severe effects, and 1 (2 percent) had a life-
`threatening adverse effect. The most common adverse
`effects of at least moderate severity were rash (17 chil-
`dren; 30 percent), diarrhea (10 children; 18 percent),
`neutropenia (7 children; 12 percent), and biochem-
`ical abnormalities (7 children; 12 percent). Most rash-
`es (88 percent) appeared within 14 days after the ini-
`tiation of study treatment (median, 9; range, 6 to 205)
`and lasted for a median of 6 days (range, 2 to 37).
`Most rashes were maculopapular and pruritic. One
`child had target lesions juxtaposed with a maculo-
`papular rash, and two children had urticarial lesions.
`None of the children with a moderate rash were fe-
`brile at the time of the rash, and none had mucous-
`membrane involvement. Severe adverse effects consist-
`ed of neutropenia in two children and hepatic toxicity,
`diarrhea and neutropenia, and rash with fever (tem-
`perature, >39¡C) in one child each. The only life-
`threatening adverse effect was neutropenia in a child
`with concurrent varicella. Central nervous system tox-
`icity was uncommon. Eight children had mild dizzi-
`ness or lightheadedness that resolved once efavirenz
`was given at bedtime rather than in the morning.
`Treatment was discontinued in 14 children. One
`child was found to be ineligible owing to prior use
`of nevirapine. Another was unable to take efavirenz
`capsules. Treatment was discontinued at the request
`of the parent, guardian, or investigator in four chil-
`dren in whom a moderate rash developed and, as spec-
`ified by the protocol, in one child who had a severe
`rash. None of these children resumed treatment af-
`ter the rash resolved. The median duration of rash
`in the children who discontinued treatment was six
`days, as compared with nine days in those who con-
`tinued treatment (P=0.14). Other reasons for discon-
`tinuation were virologic failure (six children) and
`noncompliance (one child).
`
`Pharmacokinetic Evaluations
`Pharmacokinetic evaluations of efavirenz and nel-
`finavir were performed on 50 children at weeks 2 and
`6. At week 2, the mean dose of efavirenz was 310 mg
`per day (11.7 mg per kilogram per day), and the mean
`value for the area under the curve from 0 to 24 hours
`was 218 (cid:181)mol per liter ?
`hour. One child was exclud-
`ed from the analysis because of apparently reduced
`clearance (a value of 1313 (cid:181)mol per liter ?
`hour for
`the 24-hour area under the curve). On the basis of the
`evaluations at week 2, increased doses of efavirenz
`were recommended for 22 children and decreased
`doses were recommended for 3 children. The aver-
`age dose of efavirenz was 369 mg per day (14.2 mg
`per kilogram per day), and the average value for the
`24-hour area under the curve was 244 (cid:181)mol per
`liter ?
`hour at week 6. On the basis of the evaluations
`at week 6, increased doses of efavirenz were recom-
`mended for nine children, and decreased doses were
`recommended for six children. An intention-to-treat
`
`Page 3 of 8
`
`
`
`COMBINATION THERAPY IN CHILDREN INFECTED WITH HIV-1
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Analysis of actual treatment(cid:13)
`Intention-to-treat analysis(cid:13)
`Conservative analysis
`
`100
`
`90
`
`80
`
`70
`
`60
`
`50
`
`40
`
`30
`
`20
`
`10
`
`0
`
`Percent with HIV-1 RNA Level of <400 Copies/ml
`
`0
`
`2 4
`
`8
`
`12
`
`20
`
`32
`
`40
`
`48
`
`Week
`
`Figure 1.
` Percentage of Children with Plasma HIV-1 RNA Levels of Less Than 400 Copies per Milliliter, According to an
`Analysis in Which Only the HIV-1 RNA Levels of Children Who Were Receiving Treatment Were Included (Analysis of
`Actual Treatment), an Intention-to-Treat Analysis, and an Analysis in Which HIV-1 RNA Levels Were Assigned a Value of
`More Than 400 Copies per Milliliter for All Visits after the Discontinuation of Treatment (Conservative Analysis).
`Bars indicate 95 percent confidence intervals.
`
`analysis showed that values for the 24-hour area un-
`der the curve were within the target range in 44 per-
`cent of children (22 of 50) at week 2 and 56 percent
`(28 of 50) at week 6.
`The average doses of nelfinavir were 606 mg given
`three times a day (23.6 mg per kilogram) at week
`2 and 631 mg given three times a day (24.6 mg per
`kilogram) at week 6. The corresponding mean values
`for the area under the curve from zero to eight hours
`were 21.6 mg per liter ?
`hour at week 2 and 19.9 mg
`per liter ?
`hour at week 6. On the basis of the evalua-
`tions at week 2 and week 6, the doses of nelfinavir
`were increased in two children and one child, respec-
`tively. Seventy-four percent of children (37 of 50) had
`values for the eight-hour area under the curve above
`the target range at week 2, and 80 percent (40 of
`50) had such values at week 6. On the basis of eval-
`uations performed at week 10 in 15 children, the
`doses of efavirenz were increased in 5 children and
`decreased in 1, and the dose of nelfinavir was in-
`creased in 1 child.
`
`Effect of Treatment on Plasma HIV-1 RNA Levels
`The percentage of children with plasma HIV-1
`RNA levels of less than 400 copies per milliliter rose
`rapidly after the initiation of treatment and remained
`
`stable from week 8 to week 48 in the intention-to-
`treat analysis, the analysis in which only the HIV-1
`RNA levels of children who were receiving treat-
`ment were included, and the analysis in which HIV-1
`RNA levels were assigned a value of more than 400
`copies per milliliter for all visits after the discontin-
`uation of treatment (Fig. 1). After the exclusion of
`2 children who discontinued treatment shortly after
`entry, we found that 50 of 55 children (91 percent)
`had decreases in plasma HIV-1 RNA levels to less
`than 400 copies per milliliter. Eight of these 50 chil-
`dren (16 percent) subsequently fulfilled the criteria
`for virologic failure. At week 48, the percentage of
`children with plasma HIV-1 RNA levels of less than
`400 copies per milliliter was 81 percent in the analy-
`sis in which only the HIV-1 RNA levels of children
`who were receiving treatment were included, 76 per-
`cent in the intention-to-treat analysis, and 61 per-
`cent in the analysis in which HIV-1 RNA levels were
`assigned a value of more than 400 copies per millili-
`ter for all visits after the discontinuation of treatment.
`The proportion of children with plasma HIV-1
`RNA levels of less than 50 copies per milliliter rose
`rapidly during the first 12 weeks of therapy and then
`more slowly (Fig. 2). At week 48, the percentage of
`children with plasma HIV-1 RNA levels of less than
`
`Volume 341 Number 25
`
`Æ
`
`1877
`
`Page 4 of 8
`
`
`
`The New England Journal of Medicine
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Analysis of actual treatment(cid:13)
`Intention-to-treat analysis(cid:13)
`Conservative analysis
`
`100
`
`90
`
`80
`
`70
`
`60
`
`50
`
`40
`
`30
`
`20
`
`10
`
`0
`
`Percent with HIV-1 RNA Level of <50 Copies/ml
`
`0
`
`2 4
`
`8
`
`12
`
`20
`
`32
`
`40
`
`48
`
`Week
`
`Figure 2.
` Percentage of Children with Plasma HIV-1 RNA Levels of Less Than 50 Copies per Milliliter, According to an
`Analysis in Which Only the HIV-1 RNA Levels of Children Who Were Receiving Treatment Were Included (Analysis of
`Actual Treatment), an Intention-to-Treat Analysis, and an Analysis in Which HIV-1 RNA Levels Were Assigned a Value of
`More Than 50 Copies per Milliliter for All Visits after the Discontinuation of Treatment (Conservative Analysis).
`Bars indicate 95 percent confidence intervals.
`
`50 copies per milliliter was 70 percent in the analysis
`in which only the HIV-1 RNA levels of children who
`were receiving treatment were included, 63 percent
`in the intention-to-treat analysis, and 53 percent in the
`analysis in which HIV-1 RNA levels were assigned a
`value of more than 50 copies per milliliter for all vis-
`its after the discontinuation of treatment.
`Plasma HIV-1 RNA levels fell dramatically during
`the first 20 weeks of therapy (Fig. 3). At both weeks
`32 and 48, the upper 95 percent confidence bound
`for the median change was `2.7 log copies per mil-
`liliter; thus, there is a 95 percent probability that the
`median decrease was at least this amount at these
`points.
`
`Predictors of a Decrease in Plasma HIV-1 RNA
`to Undetectable Levels
`We used Cox proportional-hazards regression mod-
`els to examine the effect of base-line characteristics
`on the likelihood that plasma HIV-1 RNA levels
`would decrease to less than 400 copies per milliliter
`and 50 copies per milliliter during treatment (Table
`2). In univariate analysis, a higher percentage of CD4
`cells at base line and a higher age- and sex-adjusted
`z score for weight significantly increased the likelihood
`that plasma HIV-1 RNA levels would drop below
`
`400 copies per milliliter, whereas only the latter var-
`iable was associated with the likelihood of a drop in
`HIV-1 RNA levels to less than 50 copies per millili-
`ter. Higher plasma HIV-1 RNA levels at base line sig-
`nificantly decreased the likelihood that plasma HIV-1
`RNA levels would become undetectable, whereas
`changing the regimen of nucleoside reverse-transcrip-
`tase inhibitors within two weeks after entry into the
`study did not have a significant effect.
`In models with multiple covariates, a high plasma
`HIV-1 RNA level at base line significantly decreased
`the likelihood that plasma HIV-1 RNA levels would
`drop below 400 copies per milliliter or 50 copies per
`milliliter, whereas the age- and sex-adjusted z score
`for weight was a significant risk factor for the failure
`of levels to drop below 400 copies per milliliter but
`not 50 copies per milliliter.
`
`Effect of Treatment on CD4 Cells
`At week 48, the CD4 cell count had increased by a
`median of 74 per cubic millimeter from base line (da-
`ta not shown), and the absolute percentage of CD4
`cells had increased by a median of 3 percent (Fig. 3).
`When all CD4 measurements during the 48-week pe-
`riod were combined, the increases in median per-
`centage and number of CD4 cells were statistically
`
`1878
`
`Æ
`
`December 16, 1999
`
`Page 5 of 8
`
`
`
`COMBINATION THERAPY IN CHILDREN INFECTED WITH HIV-1
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Median Change in HIV-1 RNA Level (log copies/ml)
`
`0.0
`
``0.5
`
``1.0
`
``1.5
`
``2.0
`
``2.5
`
``3.0
`
``3.5
`
`CD4 cells
`
`HIV-1 RNA
`
`8
`
`0246
`
``2
`
`Median Change in Percentage of CD4 Cells (%)
`
`0
`
`2 4
`
`8
`
`12
`
`20
`
`32
`
`40
`
`48
`
`Week
`
`NO. OF MEASUREMENTS
`HIV-1 RNA(cid:13)
`56(cid:13)
`52(cid:13)
`50(cid:13)
`CD4 cells
`56
`53
`50
`
`49(cid:13)
`48
`
`50(cid:13)
`50
`
`49(cid:13)
`48
`
`43(cid:13)
`44
`
`39 (cid:13)
`43
`
`43(cid:13)
`43
`
`Figure 3.
` Median Absolute Changes from Base Line in the Percentage of CD4 Cells and Plasma HIV-1 RNA Levels, According to an
`Analysis in Which Only the Values in Children Who Were Receiving Treatment Were Included.
`Bars indicate 95 percent confidence intervals. Arrows indicate an open interval.
`
`significant (global P values, 0.01 and <0.001, respec-
`tively).
`
`DISCUSSION
`Combination therapy with efavirenz, nelfinavir, and
`nucleoside reverse-transcriptase inhibitors was well
`tolerated by most HIV-infected children in this study.
`The most common treatment-related adverse effect
`of at least moderate severity was rash, which occurred
`in 30 percent of the children (17 of 57), an incidence
`similar to that with efavirenz therapy in HIV-infect-
`ed adults.
` Only five children (9 percent), all of whom
`19
`had a rash, discontinued treatment permanently be-
`cause of adverse effects. One child had severe rash
`with a fever (temperature, >39¡C) and thus fulfilled
`the criteria for the discontinuation of treatment. The
`other four children had moderate rashes and might
`have been able to continue treatment. Most cases of
`diarrhea were probably caused by nelfinavir, because
`diarrhea is a common side effect of nelfinavir thera-
`py. Central nervous system symptoms were reported
`in 51.9 percent of adults receiving efavirenz,
` but
`19
`they occurred much less frequently and were milder
`in the children in our study.
`
`The mean values for the 24-hour area under the
`curve for efavirenz at week 2 (218 (cid:181)mol per
`liter ?
`hour) and week 6 (244 (cid:181)mol per liter ?
`hour)
`compared favorably with the mean value of 184 (cid:181)mol
`per liter ?
`hour in adults receiving 600 mg of efavirenz
`per day.
` Similarly, the mean values for the eight-hour
`19
`area under the curve for nelfinavir at week 2 (21.6 mg
`hour) and week 6 (19.9 mg per liter ?
`per liter ?
`hour)
`were similar to the median value of 19 mg per
`liter ?
`hour in children who were receiving a mean
`dose of 23 mg per kilogram.
`20
`We observed potent and sustained antiviral effects
`in HIV-1—infected children who were treated with
`efavirenz, nelfinavir, and nucleoside reverse-trans-
`criptase inhibitors. Even in the conservative analysis
`in which HIV-1 RNA levels were assigned a value of
`more than 400 copies per milliliter for all visits after
`the discontinuation of treatment, 53 percent of the
`children had plasma HIV-1 RNA levels of less than
`50 copies per milliliter at week 48. Long-term fol-
`low-up will provide important information on the
`durability of this antiviral effect. There was also a
`significant but moderate increase in the number and
`percentage of CD4 cells, despite the fact that most
`
`Volume 341 Number 25
`
`Æ
`
`1879
`
`Page 6 of 8
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`The New England Journal of Medicine
`
`TABLE 2. EFFECTS OF BASE-LINE CHARACTERISTICS ON THE LIKELIHOOD THAT PLASMA
`HIV-1 RNA LEVELS WOULD BECOME UNDETECTABLE ON THE QUANTITATIVE ASSAY
`AND THE ULTRASENSITIVE ASSAY.*
`
`CHARACTERISTIC
`
`HIV-1 RNA <400 COPIES/ml
`RISK RATIO
`P
`(95% CI)
`VALUE
`
`HIV-1 RNA <50 COPIES/ml
`RISK RATIO
`P
`(95% CI)
`VALUE
`
`Univariate analysis
`Male sex
`Non-Hispanic black (vs. other
`racial or ethnic groups)
`CDC class N or A (vs. all other
`classes)
`Age (per year of age)
`CD4 cell count
`Per increase by a factor of 2
`Per increase by a factor of 10
`>500/mm3 (vs. ˙500/mm3)(cid:160)
`Percentage of CD4 cells
`Per unit increase
`>25% (vs. ˙25%)(cid:224)
`Age- and sex-adjusted z score for
`weight (per unit increase)
`Plasma HIV-1 RNA level
`Per increase by a factor of 10
`>10,000 copies/ml (vs.
`˙10,000 copies/ml)
`Change in NRTI antiretroviral
`therapy (vs. no change)
`Analysis with multiple covariates⁄
`Age- and sex-adjusted z score for
`weight (per unit increase)
`Plasma HIV-1 RNA level (per
`increase by a factor of 10)(cid:192)
`
`0.7 (0.4—1.4)
`0.9 (0.5—1.6)
`
`0.8 (0.4—1.6)
`
`1.0 (0.9—1.2)
`
`1.1 (0.9—1.5)
`1.5 (0.6—3.5)
`1.0 (0.5—2.2)
`
`1.03 (1.00—1.06)
`1.5 (0.7—3.1)
`1.4 (1.1—1.8)
`
`0.33
`0.68
`
`0.55
`
`0.40
`
`0.34
`0.34
`0.97
`
`0.04
`0.28
`0.02
`
`0.8 (0.4—1.7)
`0.9 (0.4—1.6)
`
`0.8 (0.4—1.6)
`
`1.0 (0.9—1.2)
`
`1.2 (0.9—1.6)
`1.8 (0.6—5.2)
`1.0 (0.4—2.2)
`
`0.63
`0.64
`
`0.55
`
`0.71
`
`0.28
`0.28
`0.97
`
`0.05
`1.03 (1.0—1.1)
`0.32
`1.5 (0.7—3.5)
`1.4 (1.05—1.94) 0.02
`
`0.3 (0.2—0.6)
`0.4 (0.2—0.7)
`
`<0.001
`0.002
`
`0.5 (0 3—0.8)
`0.4 (0.2—0.8)
`
`0.005
`0.01
`
`1.4 (0.7—2.6)
`
`0.30
`
`1.0 (0.5—2.0)
`
`0.95
`
`1 3 (1.0—1.8)
`
`0.04
`
`(cid:209)ƒ
`
`0 3 (0.2—0.6)
`
`<0.001
`
`0.5 (0.3—0.8)
`
`0.005
`
`*The limits of detection of the quantitative assay and the ultrasensitive assay were 400 copies per
`milliliter and 50 copies per milliliter, respectively. CI denotes confidence interval, CDC Centers for
`Disease Control and Prevention, and NRTI nucleoside reverse-transcriptase inhibitor.
`(cid:160)Similar results were obtained for CD4 cell counts of more than 700 per cubic millimeter (vs.
`˙700 per cubic millimeter) and more than 1000 per cubic millimeter (vs. ˙1000 per cubic milli-
`meter).
`(cid:224)Similar results were obtained for values of more than 30 percent CD4 cells (vs. ˙30 percent CD4
`cells).
`⁄A stepwise procedure was used. A risk factor was added to the model when it was significant at
`the 0 10 level in the forward-selection step; then the stepwise method deleted any variable that was
`not significant at the 0.05 level.
`ƒThe adjusted risk ratio was 1.3 (95 percent confidence interval, 1.0 to 1.8; P=0.06) in a model
`that included both plasma HI